Re: [PATCH v3 2/3] nvmem: sunxi-sid: add support for H3's SID controller

From: Maxime Ripard
Date: Tue Feb 07 2017 - 04:28:43 EST


On Mon, Feb 06, 2017 at 04:56:55PM +0800, Icenowy Zheng wrote:
> 06.02.2017, 16:54, "Maxime Ripard" <maxime.ripard@xxxxxxxxxxxxxxxxxx>:
> > On Thu, Feb 02, 2017 at 09:13:37PM +0800, Icenowy Zheng wrote:
> >>  The H3 SoC have a bigger SID controller, which has its direct read
> >>  address at 0x200 position in the SID block, not 0x0.
> >>
> >>  Also, H3 SID controller has some silicon bug that makes the direct read
> >>  value wrong at cold boot, add code to workaround the bug. (This bug has
> >>  already been fixed on A64 and later SoCs)
> >>
> >>  Signed-off-by: Icenowy Zheng <icenowy@xxxxxxxx>
> >>  ---
> >>  This patch is the part of [PATCH v2 1/1] that adds support for H3 SID
> >>  controller.
> >>
> >>   .../bindings/nvmem/allwinner,sunxi-sid.txt | 12 +++-
> >>   drivers/nvmem/sunxi_sid.c | 72 +++++++++++++++++++++-
> >>   2 files changed, 82 insertions(+), 2 deletions(-)
> >>
> >>  diff --git a/Documentation/devicetree/bindings/nvmem/allwinner,sunxi-sid.txt b/Documentation/devicetree/bindings/nvmem/allwinner,sunxi-sid.txt
> >>  index d543ed3f5363..9ab9e75a6351 100644
> >>  --- a/Documentation/devicetree/bindings/nvmem/allwinner,sunxi-sid.txt
> >>  +++ b/Documentation/devicetree/bindings/nvmem/allwinner,sunxi-sid.txt
> >>  @@ -1,7 +1,11 @@
> >>   Allwinner sunxi-sid
> >>
> >>   Required properties:
> >>  -- compatible: "allwinner,sun4i-a10-sid" or "allwinner,sun7i-a20-sid"
> >>  +- compatible: Should be one of the following (depending on your SoC):
> >>  + "allwinner,sun4i-a10-sid"
> >>  + "allwinner,sun7i-a20-sid"
> >>  + "allwinner,sun8i-h3-sid"
> >>  +
> >>   - reg: Should contain registers location and length
> >>
> >>   = Data cells =
> >>  @@ -19,3 +23,9 @@ Example for sun7i:
> >>                   compatible = "allwinner,sun7i-a20-sid";
> >>                   reg = <0x01c23800 0x200>
> >>           };
> >>  +
> >>  +Example for sun8i-h3:
> >>  + sid@01c14000 {
> >>  + compatible = "allwinner,sun8i-h3-sid";
> >>  + reg = <0x01c14000 0x400>;
> >>  + };
> >>  diff --git a/drivers/nvmem/sunxi_sid.c b/drivers/nvmem/sunxi_sid.c
> >>  index 69524b67007f..476a161ff23a 100644
> >>  --- a/drivers/nvmem/sunxi_sid.c
> >>  +++ b/drivers/nvmem/sunxi_sid.c
> >>  @@ -25,6 +25,16 @@
> >>   #include <linux/slab.h>
> >>   #include <linux/random.h>
> >>
> >>  +/* Registers and special values for doing register-based SID readout on H3 */
> >>  +#define SUN8I_SID_PRCTL 0x40
> >>  +#define SUN8I_SID_RDKEY 0x60
> >>  +
> >>  +#define SUN8I_SID_OP_LOCK 0xAC
> >>  +#define SUN8I_SID_OFFSET_MASK 0x1FF
> >>  +#define SUN8I_SID_OFFSET_SHIFT 16
> >>  +#define SUN8I_SID_LOCK_SHIFT 8
> >>  +#define SUN8I_SID_READ BIT(1)
> >>  +
> >>   static struct nvmem_config econfig = {
> >>           .name = "sunxi-sid",
> >>           .read_only = true,
> >>  @@ -34,11 +44,14 @@ static struct nvmem_config econfig = {
> >>   };
> >>
> >>   struct sunxi_sid_cfg {
> >>  + u32 value_offset;
> >>           u32 size;
> >>  + bool need_register_readout;
> >>   };
> >>
> >>   struct sunxi_sid {
> >>           void __iomem *base;
> >>  + u32 value_offset;
> >>   };
> >>
> >>   /* We read the entire key, due to a 32 bit read alignment requirement. Since we
> >>  @@ -51,7 +64,8 @@ static u8 sunxi_sid_read_byte(const struct sunxi_sid *sid,
> >>   {
> >>           u32 sid_key;
> >>
> >>  - sid_key = ioread32be(sid->base + round_down(offset, 4));
> >>  + sid_key = ioread32be(sid->base + sid->value_offset +
> >>  + round_down(offset, 4));
> >
> > This would probably be more logical to have this in sunxi_sid_read.
>
> But it's here which really access the memory...

This function is made to read a single register. What you want is to
offset all reads, and all the reads are made in sunxi_sid_read.

> >>  + if (reg_val & SUN8I_SID_READ)
> >>  + return -EIO;
> >>  +
> >>  + if (out)
> >>  + *out = readl(sid->base + SUN8I_SID_RDKEY);
> >
> > Why do you need that out parameter?
>
> The read operation by registers can really return a value --
> in fact, the fix to the pre-read value is a side effect.

Yet, you're not using it at all, so this is dead code.

Maxime

--
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

Attachment: signature.asc
Description: PGP signature